xAI Launches Grok Build 0.1: Coding Model with 256K Context for Agentic Workflows

xAI has released Grok Build 0.1, a coding-specialized language model with a 256K token context window and no text output limit, according to the company.

Model Specifications

Grok Build 0.1 accepts both text and image inputs while generating text output. The model is currently in early access and available through OpenRouter's API platform.

Key specifications:

• Context window: 256K tokens
• Output limit: None
• Modalities: Text and image input, text output
• Release date: May 20, 2025 (according to OpenRouter listing)
• Pricing: Not yet disclosed

Purpose-Built for Coding Agents

xAI claims the model is "trained specifically for agentic software engineering workflows." The company positions it as optimized for:

• Interactive coding agents
• Tool use and function calling
• Multi-step development tasks
• Long-horizon coding projects
• Automation workflows

The model powers xAI's Grok Build CLI, a command-line interface tool for developers.

Technical Context

The 256K context window places Grok Build 0.1 in the upper tier of commercially available models, matching capabilities from Anthropic's Claude 3 series and Google's Gemini 1.5 models. The unlimited output generation is notable for coding use cases where generating complete files or large code blocks is common.

xAI has not disclosed benchmark scores, parameter count, or training data details. The company also has not specified whether this model is a variant of its existing Grok-2 architecture or represents a new model family.

What This Means

Grok Build 0.1 represents xAI's first model explicitly targeting the coding assistant and agentic development market, competing directly with OpenAI's o1 models, Anthropic's Claude family, and Google's Gemini for Developers. The emphasis on "agentic workflows" and CLI integration suggests xAI is pursuing the emerging category of autonomous coding agents rather than traditional code completion. Without pricing information or benchmark data, it remains unclear how Grok Build 0.1 compares to existing alternatives in performance or cost-effectiveness. The early access designation indicates limited availability as xAI likely refines the model based on developer feedback.